Ben Kochie
Ben Kochie
Yea, I'm not sure we want to do it this way. The mountpoint, while useful, is going to lead to a lot of duplicate cardinality for some users. For example,...
Very possible to have weird data coming from the ethtool syscall has problems. The data is vendor driver dependent, so you never know what they're going to spit out. We...
Can you run a debugging exporter and scrape it manually with curl to see what it does? Something like this: ``` node_exporter \ --log.level=debug \ --collector.ethtool \ --collector.ethtool.device-exclude="^(brq|tap|veth|vxlan|virbr|usb).*$" \ --collector.disable-defaults...
The output of `ethtool -S ` would be useful.
Very strange, it doesn't differ in those files. I think the only thing we can try to do is apply the same sanitizer rules to the help text.
Please be sure to follow our bug reporting template. You have left out what kernel version and OS version you are using.
We allow CGO for Darwin, since there's a bunch of stuff that require it.
The exporter only reports what the kernel tells it to. So this is likely a bug in your kernel version.
Good find. I'm not sure how easy it will be to actually work around that problem, other than exposing the individual per-cpu counters directly so resets can be handled.
I was just reviewing this, and the original procfs PR. This is not a metric we can accept, as it is a computation against two metrics we already expose. ```...